We are looking for ways to improve the response times of the ARSDOC command.
None of the tables are particularly large, the biggest is only 2.2M rows. But our measured response times when querying a Folder are 1.2 to 2 seconds.
In the Database section of System I Navigator and IBM Navigator for I, you can look at an Advised Indexes report for the OnDemand tables. It suggests the creation of additional indexes based on certain key fields. Is this advisable?
You can also automatically create the indexes right from there but doing so could cause some issues down the road. First, the names generated are very goofy so by default it’s hard to track down what they are. Secondly, there are source control issues and locks placed on the underlying tables. If we were to upgrade OnDemand at any point, those files wouldn’t be able to be manipulated unless all the indexes were deleted first.
If you have any other suggestions for improving the performance, we would be happy to hear them.
Thanks in advance.